La PREUVE qu'Internet va MOURIR

La PREUVE qu'Internet va MOURIR

by Christophe Pauly

Humanities, Social Sciences & Thought AI & Society UBDigital and Information technology: general topicsUBJDigital and information technologies: social and ethical aspects
🎙 Christophe Pauly 👥 254K 📅 November 23, 2025 ⏱ 32 min 👁 107K 📄 expert opinion 🧭 2026-08-02
Available in: English (current) Français

Keywords

deepfakeGANdiffusion modelsinternet trustAI-generated content

Summary

The video explores the growing difficulty of distinguishing real from AI-generated content on the internet, arguing that this could lead to a collapse of trust and the ‘death’ of the internet as a reliable source of information. It traces the history of generative AI from the invention of GANs in 2014 to modern diffusion models and video generation tools like Sora and VEO. The narrator uses the example of his own potential AI identity to illustrate the pervasive suspicion. He explains the technical mechanisms of GANs and diffusion models in an accessible way, highlighting how they have evolved to produce hyper-realistic images and videos. The video discusses the psychological concept of apophenia, where humans see patterns and meaning in random data, which contributes to the suspicion. It also touches on the societal implications, such as the use of deepfakes in misinformation and the challenge of verifying evidence. The conclusion suggests that we have already passed a point of no return, and the internet is becoming a place where truth is increasingly ambiguous. The video is sponsored by Mammouth AI, a platform aggregating multiple AI models, which is disclosed at the beginning.

191 words

Critical Evaluation

The video provides a compelling and well-illustrated overview of the evolution of generative AI and its impact on trust in digital content. The narrative is engaging, using historical milestones and concrete examples to explain complex concepts like GANs and diffusion models. The explanation of GANs as a competitive game between a generator and a discriminator is particularly effective for a general audience, and the analogy of diffusion models as reconstructing a statue from dust is vivid and memorable. The video successfully raises important questions about the authenticity of visual evidence and the psychological mechanisms that drive suspicion, such as apophenia. However, the argumentation is somewhat one-sided, focusing heavily on the dangers of AI-generated content without adequately discussing countermeasures like watermarking, detection algorithms, or legal frameworks. The scientific rigor is moderate: while the video references a scientific article on deepfakes, it does not delve into the technical limitations or the current state of detection research. The sponsor segment is clearly separated and does not unduly influence the content, but it does take up a portion of the video. The title, while attention-grabbing, is somewhat hyperbolic, but the content does support the thesis that the internet is facing a crisis of trust. Overall, the video is a valuable contribution to public understanding of AI’s societal implications, but it could benefit from a more balanced perspective and deeper engagement with counterarguments.

228 words

Title / Content Match

The title is somewhat sensationalist but accurately reflects the video's central thesis about the erosion of trust on the internet due to AI-generated content.

Quality & Reliability

7/10

The video is a well-structured expert opinion piece that traces the evolution of generative AI and its societal implications. It references key technologies (GANs, diffusion models) and includes a scientific article link, but relies heavily on narrative and lacks deep technical detail. The presence of a sponsor segment is disclosed but does not affect the score.

Chapters

Cited Sources

  • Determining authenticity of video evidence in the age of artificial intelligence and in the wake of Deepfake videos — Referenced as a scientific article read during research, likely used to support claims about deepfake challenges.
  • Deepfake : L'IA au service du faux — Recommended as a book on deepfakes, providing further reading.
  • Interview with a scientist on AI surpassing humans — Recommended as an interview with a scientist, likely to provide additional perspectives.

Concurring Sources

  • Determining authenticity of video evidence in the age of artificial intelligence and in the wake of Deepfake videos — Supports the video's claims about the challenges of verifying video evidence in the age of AI.

External References

Contribution & Novelties

The video offers a clear and accessible synthesis of the evolution of generative AI, from GANs to diffusion models, and its societal implications. It uniquely frames the issue as a crisis of trust, linking psychological concepts like apophenia to the current suspicion of AI-generated content. The narrative is engaging and well-structured, making complex technologies understandable to a broad audience.

Pour aller plus loin :

  • Generative adversarial network — Wikipedia article on GANs, foundational to the video’s explanation.
  • Diffusion model — Wikipedia article on diffusion models, the technology behind modern image generators.
  • Deepfake — Wikipedia article on deepfakes, directly relevant to the video’s topic.
  • Apophenia — Wikipedia article on the psychological phenomenon discussed in the video.

115 words

Radar Profile

The radar profile shows a balanced performance with high scores in quantity of information and moderate scores in technical level and reliability. The video excels in providing a comprehensive overview but could improve in technical depth and source diversity.

Reliability 7/10

💬 Positif : Sur les 30 commentaires analysés, le climat est très positif, avec des éloges sur la qualité du contenu et la réflexion, bien que certains expriment une fatigue face aux vidéos générées par IA.